Skip to content

Clinical workflow

California SB 1120: Defending Notes Against AI Denials

SB 1120 compliance guide: structure clinical notes to defend against payer AI denials ahead of the 2027 enforcement deadline.

Book a demo✦
10 min read
Compliance, AI Scribe, AB 3030
COMPLIANCEAUDIT-READYDISCLOSUREATTESTATION

California SB 1120 Clinical Documentation Defense and Payer AI Denial Mitigation: The 2027 Playbook

Merry AI · Thoughtfully curated clinical briefs.

Clinically reviewed by Dr. Marcus Vance, MD (Medical Informatics Advisor) · Coding review by Kaif R. (Clinical Informatics Specialist) · Sources: CY 2026 CMS PFS, NCCI Policy Manual.

California SB 1120 prohibits health plans and disability insurers from denying, delaying, or modifying care based on medical necessity unless a licensed physician or qualified health care professional makes that decision. Any AI tool used in utilization review must also base its output on the individual patient's medical history and clinical circumstances. Your strongest defense against an AI-driven denial is a note that documents those individual circumstances explicitly, in structured, coded, and dated form.

This playbook is for outpatient specialty physicians in California, with cardiology as the worked example. It covers the statutory requirements, documentation patterns that survive algorithmic review, Medical Decision Making (MDM) alignment, CPT G2211 capture, and the federal FHIR prior authorization deadline that reshapes payer adjudication in 2027.

SB 1120 requires a licensed physician, not an algorithm, to deny medical necessity

SB 1120, the Physicians Make Decisions Act, took effect January 1, 2025. It amended Health and Safety Code Section 1367.01, which governs plans regulated by the Department of Managed Health Care (DMHC). It also amended Insurance Code Section 10123.135, which governs insurers regulated by the California Department of Insurance (CDI).

The statute imposes several conditions on any artificial intelligence, algorithm, or software tool a payer uses for utilization review or utilization management:

  • Individualized basis required: determinations must rest on the enrollee's medical history, the individual clinical circumstances presented by the requesting provider, and other relevant information in the clinical record.
  • No group-only decisions: the tool cannot base its determination solely on a group dataset.
  • No supplanting clinicians: the tool cannot replace health care provider decision-making.
  • Nondiscrimination and fair application: the tool must not discriminate and must be applied fairly and equitably.
  • Open to regulator inspection: the tool and its written policies must be available for audit and compliance review by DMHC or CDI.
  • Periodic performance review: payers must periodically review and revise the tool to maximize accuracy and reliability.
  • Purpose-limited patient data: patient data cannot be used beyond its intended and stated purpose.

The core clinical protection is the human decision requirement. A denial, delay, or modification based on medical necessity must be made by a licensed physician or a qualified health care professional who is competent to evaluate the specific clinical issues. That reviewer must consider the requesting provider's recommendation, the enrollee's medical history, and the individual clinical circumstances. For the specialty-specific defense layer, see the California SB 1120 Utilization Review Defense brief.

Key point for physicians: SB 1120 regulates payers. Your note supplies the evidence it requires.

Payer AI must weigh individual circumstances, so your note must state them

An algorithm can only weigh what it can read. When a note says "HF, stable, continue meds," an automated scorer sees a population-level diagnosis with no individual modifiers. That is the exact failure mode SB 1120 prohibits, but the payer can still argue the record contained nothing individualized to weigh.

Defensible documentation closes that gap by making five elements machine-readable and human-verifiable:

  1. Specific, coded diagnoses: the most specific ICD-10-CM code supported, never an unspecified default when type and acuity are known.
  2. Dated objective values: imaging, labs, and functional measures with dates, so trend is visible.
  3. Prior therapy history: what was tried, at what dose, for how long, and why it failed or was contraindicated.
  4. Risk-modifying comorbidities: conditions that change the risk-benefit calculation for this patient specifically.
  5. Explicit decision reasoning: one or two sentences linking findings to the ordered service.

The fifth element matters most. Reviewers, human or automated, consistently overturn on reasoning they can quote. A sentence such as "Proceeding to cardiac MRI because echo windows are nondiagnostic and ischemic etiology must be excluded before ICD referral" is quotable evidence of individual circumstances.

A defensible heart failure note pairs I50.22 or I42.0 with trend data

Heart failure and cardiomyopathy generate a high volume of prior authorization requests for advanced imaging, device evaluation, and branded guideline-directed medical therapy. Code specificity drives the first algorithmic pass.

Use I50.22 (ICD-10-CM) for chronic systolic (congestive) heart failure when the patient has reduced ejection fraction and a chronic course. Document the most recent LVEF with date and modality, NYHA functional class, and the natriuretic peptide trend. When decompensation is present, the acuity code changes, and the note must say why.

Use I42.0 (ICD-10-CM) for dilated cardiomyopathy. Pair it with the heart failure code when both apply, following ICD-10-CM sequencing guidance. Document chamber dimensions, the etiologic workup completed, and family history when genetic evaluation is in play.

A compact defensible assessment reads like this: "Chronic HFrEF (I50.22) due to dilated cardiomyopathy (I42.0). LVEF 30% on TTE 03/2026, down from 38% in 09/2025. NYHA III. NT-proBNP rising over two draws. Intolerant of sacubitril/valsartan uptitration due to symptomatic hypotension. Ordering cardiac MRI to assess for scar before ICD candidacy decision."

Specificity beats volume: one dated trend outweighs a paragraph of narrative.

MDM level, clinical indicators, and audit defenses must align in every note

The 2021 AMA office E/M framework, which CMS continues to apply under the CY 2026 Physician Fee Schedule, scores level by MDM or total time. Payer AI tools frequently flag a mismatch between the billed level and the documented problem, data, and risk elements. The table below maps common cardiology scenarios to defensible documentation.

Clinical scenarioKey clinical indicators to documentMDM threshold (2 of 3 elements)Typical levelAudit and SB 1120 defense
Two stable chronic conditions (e.g., I50.22 plus hypertension)LVEF with date, NYHA class, BP trend, current GDMT dosesModerate problems (2+ stable chronic illnesses); prescription drug management99214Medication decisions stated explicitly, not "continue meds"
Chronic HF with mild exacerbationWeight gain, edema grade, NT-proBNP change, diuretic adjustmentModerate problem (chronic illness with exacerbation); prescription drug management99214Dated objective change proves exacerbation, not stability
Dilated cardiomyopathy (I42.0) with declining LVEFSerial LVEF, chamber dimensions, workup results, device candidacy reasoningModerate to high problem; independent interpretation of imaging; moderate risk99214 or 99215Independent interpretation documented as your own read
Severe HF exacerbation, hospitalization consideredRest dyspnea, hypoxia, renal function, decision on admissionHigh problem (severe exacerbation); high risk (decision regarding hospitalization)99215Admission decision documented even if patient declines
Drug therapy requiring intensive toxicity monitoring (e.g., amiodarone, dofetilide)Thyroid, liver, pulmonary, and QTc monitoring values with datesHigh risk (drug therapy requiring intensive monitoring for toxicity)99215 when problem and data also supportMonitoring results named, dated, and acted upon

The audit defense column reflects a single principle: every MDM element must be traceable to a sentence in the note. Total time is an alternative pathway, but time-based billing does not substitute for individualized clinical reasoning in a medical necessity dispute.

CPT G2211 complexity capture recovers $15,600 or more annually

HCPCS add-on code G2211 recognizes the inherent complexity of office and outpatient E/M visits that serve as the continuing focal point for a patient's care, or that provide ongoing care for a single serious or complex condition. Outpatient cardiologists managing longitudinal heart failure meet the second prong routinely, yet many practices underbill it.

At roughly $16 per eligible Medicare visit under the CY 2026 PFS, capturing G2211 on about 975 qualifying visits per year (roughly 20 per week) recovers $15,600+ in annual revenue. The add-on requires no extra visit time, only documentation of the longitudinal relationship.

Document the longitudinal relationship explicitly. A sentence such as "I am managing this patient's chronic HFrEF longitudinally, including GDMT titration and device candidacy" establishes the qualifying relationship. Since CY 2025, CMS permits G2211 with a modifier 25 E/M only when the other same-day service is an annual wellness visit, vaccine administration, or a Medicare Part B preventive service. Review the NCCI Policy Manual for modifier 25 rules before appending.

G2211 requires documented continuity, not added time or a separate problem.

Merry AI Pro costs 1.3% of a $48,000 fully loaded MA

The documentation burden described above is real. Many practices respond by adding a medical assistant or scribe at a fully loaded cost of about $48,000 per year once wages, benefits, payroll taxes, and training are included.

Merry AI Pro costs $54 per month on annual billing, or $648 per year. Against the $48,000 denominator, that is 1.3% of the labor cost. Merry AI drafts structured notes that surface dated values, coded diagnoses, prior therapy history, and explicit decision reasoning, the same elements SB 1120 requires payers to weigh.

The physician remains the author. Every draft requires clinician review and attestation before signing. That review step is both a clinical safety control and a compliance requirement under state rules tracked in the State AI Scribe Rules & Statutory Compliance Hub.

CMS-0057-F requires payer FHIR prior authorization APIs by January 1, 2027

The CMS Interoperability and Prior Authorization Final Rule (CMS-0057-F) applies to Medicare Advantage organizations, state Medicaid and CHIP programs, Medicaid managed care plans, and qualified health plan issuers on the federal exchanges. It does not directly govern state-regulated commercial plans, which remain under SB 1120 and Section 1367.01.

The rule sets staged deadlines that affect 2026 and 2027 denial patterns:

  • Beginning January 1, 2026: impacted payers must issue decisions within 72 hours for expedited requests and 7 calendar days for standard requests, and must provide a specific reason for denials.
  • Beginning March 31, 2026: impacted payers must publicly report prior authorization metrics, including approval and denial rates.
  • By January 1, 2027: impacted payers must implement a FHIR-based Prior Authorization API, along with Provider Access and Payer-to-Payer APIs.

Structured documentation gains leverage in 2027. When payer systems ingest FHIR resources directly, coded diagnoses, dated observations, and medication histories flow into adjudication without manual abstraction. Unstructured narrative notes will stall or trigger requests for additional information. Medicare Advantage plans must also continue to base medical necessity decisions on individual circumstances under CMS-4201-F, a federal parallel to SB 1120.

2027 adjudication rewards structure: coded, dated data moves faster than narrative.

A five-step workflow converts SB 1120 rights into overturned denials

Use this workflow when a commercial plan regulated by DMHC or CDI denies, delays, or modifies a service on medical necessity grounds:

  1. Read the denial rationale: identify whether it cites criteria, a dataset, or an individual clinical finding. A rationale that ignores documented individual circumstances is your opening.
  2. Request the reviewer's credentials: SB 1120 requires a licensed physician or qualified professional competent in the specific clinical issue. Ask for specialty during peer-to-peer.
  3. Quote your own note: cite the dated values, failed therapies, and decision reasoning verbatim. Point to where the record contained the individual circumstances the plan was required to consider.
  4. Track statutory timelines: Section 1367.01 requires prior authorization decisions within 5 business days for routine requests and 72 hours when the enrollee faces imminent and serious threat to health.
  5. Escalate through the regulator: if internal appeal fails, the enrollee may file a grievance and request Independent Medical Review through DMHC or CDI. Your documentation becomes the IMR record.

Practices that log every denial by payer, service, and rationale build pattern evidence. Repeated group-dataset rationales across unrelated patients suggest algorithmic determinations that regulators can investigate.

AB 3030 governs AI-generated patient communications, not payer review

AB 3030 is a separate California law, effective January 1, 2025, that applies to health facilities, clinics, and physician offices using generative AI to create patient communications about clinical information. Those communications must include a disclaimer stating they were AI-generated and instructions for contacting a human provider.

The disclosure requirement does not apply when a licensed or certified health care provider reads and reviews the communication before it is sent. Full criteria are outlined in California AB 3030 Generative AI Disclosure Requirements. For SB 1120 defense, the relevant point is that physician-reviewed, attested documentation carries clinical authority in both the patient-facing and payer-facing record.

Frequently asked questions about California SB 1120 documentation defense

Does SB 1120 apply to AI scribes used by physicians?

No. SB 1120 regulates health care service plans and disability insurers that use artificial intelligence for utilization review or utilization management. Physician-side ambient scribes fall under other rules, including AB 3030 patient communication disclosures, but documentation they produce becomes the evidentiary record payers must individually review.

When did SB 1120 take effect?

SB 1120 took effect January 1, 2025, amending Health and Safety Code Section 1367.01 and Insurance Code Section 10123.135. Enforcement runs through the Department of Managed Health Care and the California Department of Insurance, both of which may audit payer algorithms and review medical necessity denial files.

What documentation best defends against an AI-driven denial?

Document the individual patient, not the diagnosis category. Include dated objective values such as ejection fraction, NYHA class, and BNP trend, failed or contraindicated prior therapies, comorbidities altering risk, and the explicit reasoning linking findings to the order. SB 1120 requires reviewers to weigh exactly these circumstances.

Can I challenge the reviewer's qualifications under SB 1120?

Yes, in practice. SB 1120 requires that any denial, delay, or modification based on medical necessity be made by a licensed physician or qualified health care professional competent in the clinical issue. Requesting the reviewer's specialty and credentials during peer-to-peer discussions tests that requirement directly and builds your grievance record.

How does the 2027 FHIR Prior Authorization API change denial mitigation?

Under CMS-0057-F, Medicare Advantage, Medicaid, CHIP, and federal exchange plans must deploy a FHIR Prior Authorization API by January 1, 2027. Structured documentation that maps cleanly to coded fields such as I50.22 and I42.0 will move through automated adjudication faster and expose weak, unstructured notes sooner.

Merry AI TeamClinical Intelligence Team
10 min read